LLM–powered transcription and implementation of audiovisual radiology reports in elite soccer medicine: A pilot study and two-year clinical impact review
Year: 2025
Type: article
Abstract: Background Musculoskeletal injuries are a major burden in elite soccer, with direct consequences on team performance and athlete health. Traditional radiology reporting—largely text-based—can hinder clarity, especially in cases involving complex or multilayered injuries. The purpose of the study is to evaluate the integration of GPT-4–powered transcription and translation of audiovisual (AV) radiology video reports in elite soccer medicine and its impact on clinical communication, workflow, and decision-making. Materials and Methods A dual-phase study was conducted: (1) a pilot implementation of AV reports for elite soccer athletes, integrating GPT-4 automatic transcription, multilingual translation, and EMR embedding; and (2) a two-year retrospective analysis of 540 AV reports with survey-based evaluation from 36 clinicians. Outcomes assessed included comprehension, workflow impact, interdisciplinary communication, and user experience. Clinical and imaging data were anonymized before analysis and a written informed consent was obtained from all participants. Results GPT-4 enhanced diagnostic communication by producing simultaneous professional and layperson summaries in multiple languages. Among 36 survey respondents (response rate: 80%), 90% reported improved diagnostic understanding, 68% reported faster clinical decisions, and 79% reported improved interdisciplinary communication. AV reports were accessed within 24hours in 87% of cases. Challenges included full video length (41%) and EMR integration (29%). Conclusion GPT-4–powered AV radiology reports represent a promising innovation in elite sports medicine. Future development should emphasize EMR interoperability, concise summaries, and broader language support to optimize clinical utility.
Authors: Ivan Rodrigues Barros Godoy, Tatiane Cantarelli Rodrigues, Luís Pecci Neto, Abdalla Skaf
Institutions: Hospital do Coração, Universidade Federal de São Paulo
Language: English
Subfield: Health Informatics
Field: Medicine
Domain: Health Sciences
Open Access status: gold
Estimated APC paid: $2,900
OpenAlex ID: W7114995559